README
Equilibrium Effects of Health Care Price Information
Zach Y. Brown (University of Michigan, Department of Economics)


////////////////////////////////////
                DATA
////////////////////////////////////

The primary dataset used in this project is from the New Hampshire Comprehensive Health Care Information System (CHIS). This limited-use dataset contains sensitive patient information. The data use agreement does not allow for publicly posting the data. However, interested researchers can obtain the datasets through the New Hampshire Department of Health and Human Services. The online documentation and application instructions can be found at https://nhchis.com. The website also has a full data dictionary. The primary datasets are individual claims with provider and individual identifies and monthly individual membership in insurance plans. The NH DHHS can also provide website traffic data. For additional information, one can also contact:

Technical Support for Data Submissions:
Email: NHCHISSupport@Milliman.com

New Hampshire State Contacts:
Mary Fields, New Hampshire Department of Health and Human Services
Email: Mary.Fields@dhhs.state.nh.us

Maureen Mustard, New Hampshire Insurance Department
Email: Maureen.Mustard@ins.nh.gov 

This project also makes use of publicly available data. CMS data is used to obtain procedure descriptions, ACS data is used to obtain zip code characteristics, and the Medical Care Services CPI is used to inflation adjust prices. These data have been included in the "Data" folder.

////////////////////////////////////
               PROGRAMS
////////////////////////////////////

All programs were run using State 13.0 on Mac OS-X.

IMPORT
01_1_convert_provider_detail.do - Cleans information about the providers
01_2_convert_ref_tables.do - Cleans reference tables that help interpret codes
01_3_convert_mm.do - Imports a member month file. All years can be imported by running run_import_all_years.sh
01_4_convert_med.do - Imports a medical claims file. All years can be imported by running run_import_all_years.sh
02_1_import_membermonth.do - Cleans and combines member month files.
02_2_import_med_clm.do - Cleans and combines medical claims.
03_import_zip_characteristics.do - Imports zip code characteristics. Raw data can be downloaded from the Census website.
import_cms_cpt_codes.do - Imports CPT codes with descriptions from CMS
import_nhhealthcost_procedures.do - Imports list of procedures on healthcost website
import_price_index.do - Imports CPI Health Care Services index from BLS
import_zip_to_county.do - Imports crosswalk linking zip codes and county

BUILD
01_make_proc_id.do - Make a unique identifier for procedure codes that is consistent across datasets
02_build_proc_info.do - Gets list of procedures
03_build_deductible_info.do - Finds each individuals deductible status
04_build_payer_info.do - Cleans payer info, including classifying insurers
05_build_provider_info.do - Classifies providers
06_build_membership_info.do - Cleans membership information, including each individual's insurance enrollment by month
08_make_prov_cw_crosswalk.do - Makes provider crosswalk linking provider IDs that are specific to insurers to universal provider IDs
09_clean_cms_cpt_codes.do - Finds procedures that are on website and creates separate procedure IDs when descriptions change
10_build_radiology_visits.do - Merges relevant datasets and creates visit level prices
11_calc_distance.do - Calculate distance from each patient to provider
12_make_charlson_index.do - Creates Charlson score for each patient. This program is memory intensive (may requires >50GB of memory or it will crash)
13_merge_stuff.do - Takes main visit level dataset and merges on patient membership info, zip code characteristics, provider crosswalk, etc
14_final_visit_prep.do - Subsets final dataset for analysis and creates relevant dependent variables.
15_build_hhi.do - Calculates county HHI
16_find_price_ptiles_by_cnty.do - Finds whether a provider's price is in lowest decile for the county-year-procedure-payer. Used for examining demand-side effects.

ANALYSIS
01_potential_savings.do - Makes Table 1
02_website_traffic_chart.do - Makes Figure 1
03_website_proc_availability.do - Makes Table 2
04_indiv_summary.do - Makes Table 3
05_visits_summary.do - Makes Table 4
06_base_effect_charts.do - Makes Figure 2
07_base_regs.do - Makes Table 5
08_stratify_regs.do - Makes Table 6
09_search_regs.do - Makes Table 7
10_supply_side_charts_byhalfy.do - Makes Figure 3
11_supply_side_regs.do - Makes Table 8
12_price_dispersion_regs.do - Makes Table 9

APPENDIX
A01_insurer_trends_chart.do - Makes Figure A-1
A02_robustness_regs.do - Makes Table A-2
A03_placebo_regs.do - Makes Table A-3
A04_supply_side_regs_robustness.do - Makes Table A-4
A05_insurer_event_study.do - Makes Table A-5
A06_utilization_regs.do - Makes Table A-6



